loading...
Sonia targets the vast majority priced out of mental health care – with on-demand sessions via text or voice at a fraction of clinical cost.
ENTRY ANGLES
AI-powered therapy applications · AI therapists targeting affordability-constrained users · AI therapy in meditation and mindfulness segment
VERTICALS
CAPABILITIES
AI/LLM technology for therapeutic conversation, Mental health domain expertise, Ability to reach price-sensitive or privacy-conscious users
Sonia wants to make mental health support accessible to everyone.
The app lets any user book a therapy session with an AI assistant playing the role of a psychologist. Sessions typically run about 30 minutes, and you can engage via text or voice.
The advantages over traditional therapy are structural:
- Affordable. A live therapist session in the US costs $100–200 on average. Sonia claims to cost "1% of that." There's also a lower psychological barrier – many people find it easier to open up to an AI than to another human.
- Accessible. More than 100 million Americans live in areas where mental health professionals are scarce – roughly one therapist per 30,000 residents. Getting an appointment when you actually need one is often impossible.
- Available 24/7. You can use the app at 2 a.m. when anxiety is keeping you awake. A therapist's office cannot accommodate that.
The startup recently added a shorter-format option: a five-minute check-in for moments of acute stress or anxiety, designed to decompress quickly rather than work through anything deep.
In parallel, Sonia is building a B2B offering – selling to companies that want to give employees access to AI mental health support, with the goal of reducing stress, limiting burnout, and improving overall productivity.
Sonia is currently in the Y Combinator batch and has received its initial $500,000 from the accelerator.
Mental health is a substantial and underserved problem.
As of 2022, around 20% of American adults – roughly 50 million people – experience some form of mental health condition. A quarter of those face serious challenges.
At $100–200 per session, and with multiple sessions typically required, meaningful treatment is out of reach financially for a large portion of the people who need it.
That supply-demand mismatch is fueling market growth. The US therapy market was valued at $92 billion in 2023 and is projected to reach $144 billion by 2032. Online therapy is growing faster than the overall mental health sector – and replacing the human at the other end of the internet with an AI is the logical next step.
In most cases, patients probably won't notice a dramatic difference. A lot of what people need from therapy is the ability to articulate what they're feeling and receive support and encouragement. The most common therapeutic techniques are not especially complex – and a well-designed AI can handle the 20% of methods that account for 80% of outcomes.
This trend is worth tracking. A [recent review](/review/milliard-svetit-na-vzljote) covered Maia – an AI therapist built specifically for couples. Maia is also currently in Y Combinator. That's not a coincidence – it reflects a real moment: founders see the opportunity opening up, and Y Combinator is willing to bet on it.
The mental health app market was valued at $5.2 billion in 2022, with projections pointing to $23.8 billion by 2032 under organic growth assumptions.
AI therapists will likely accelerate that significantly – for two reasons:
- Money that currently flows to human therapists will begin shifting to AI-powered alternatives.
- AI therapy dramatically widens the addressable market, reaching people who can't afford traditional sessions or who aren't comfortable discussing their problems with another person.
The broader mental health services market sits at $147.2 billion in 2023, growing to $227.5 billion by 2032. If AI applications capture even 10% of that – a conservative assumption – the mental health app market could roughly double its projected 2032 size.
The opportunity: build AI therapy apps. The meditation and mindfulness segment alone has already produced a handful of companies valued at over $1 billion. An AI that can do real therapeutic work has an even larger ceiling – for anyone who gets in early enough.